Recent studies showed that hemogenic endothelial cells in the aortic-gonado-mesonephros (AGM) region of the developing mouse embryo give rise to most fetal liver and bone marrow hematopoietic cells. In addition, the earliest step of hematopoiesis, the emergence of hematopoietic stem cells (HSC) from the hemogenic endothelium, is controlled by Runx1. However, it is not clear whether Runx1 is required for the emergence of all definitive HSCs, both in the embryos and in the adult bone marrow. Through large scale DNA sequencing of ENU-treated zebrafish, we developed a zebrafish line with a truncation mutation (W84X) in the runx1 gene. Characterization of this line showed that fetal (or larval in zebrafish) and adult hematopoiesis share a common stem cell pool, but runx1 is only required for the initiation of larval hematopoiesis. The embryos of the runx1W84X/W84X fish had complete block of larval (equivalent to fetal liver) hematopoiesis and 80% of the runx1W84X/W84X larvae died due to their bloodless phenotype. Strikingly, 20% runx1W84X/W84X larvae resumed blood circulation and grew up to fertile adults with multi-lineage adult hematopoiesis. We found that hematopoietic stem cells were retained in the runx1W84X/W84X embryos and were able to initiate adult hematopoiesis in the absence of runx1. These findings suggest that even though fetal and adult hematopoiesis share common precursors, their initiation requires distinct regulatory programs (Sood et al., Blood 2010). In collaboration with NIH Chemical Genomics Center we have conducted a screen of roughly 1 quarter million chemicals for inhibitors of RUNX1-CBFbeta interaction, which has been demonstrated to be a critical interaction for several types of human acute myeloid leukemia. Over 130 candidate compounds have been identified after the initial screen and subsequent confirmation/validation experiments. We have developed a zebrafish-based assay to determine if the candidate compounds block RUNX1-CBFbeta interaction in vivo. The assay is to incubate transgenic zebrafish embryos with GFP expression in various blood cell types (such as cd41-GFP and lck-GFP for thrombocytes and lymphocytes, respectively) with the compounds and observe for hematopoietic defects as reflected by the number of GFP+ cells. The GFP+ cells are expected to decrease in the embryos if Runx1 function is blocked by the candidate compounds, since Runx1 is required for definitive hematopoiesis. Through this approach we have identified three structurally related compounds that were able to reduce circulating GFP+ cells in the embryos, which essentially phenocopied the runx1 mutant embryos, while otherwise the embryos looked normal. One of the three compounds has been administered to our mouse leukemia model and showed that it could reduce leukemia burden in the mice, validating our approach to pre-screen the candidate compounds in the zebrafish model. In addition, we are using these compounds to determine the temporal requirement of runx1 during zebrafish embryonic hematopoiesis.